The goal of this application is to develop a chemo genomic analysis methodology for reconstructing signaling and transcription networks in bacteria. In particular, Escherichia coli responses to reactive nitrogen oxide species (RNOS), such as nitric oxide and S-nitrosothiols, will be used as a model system. Instead of relying solely on association relationships, which does not reveal chemical details of signaling mechanisms, the proposed research will develop a paradigm to incorporate chemical reactivity information in the transcriptomic analysis of E. coli. The resulting network will be mechanistically feasible and consistent with gene expression data. The proposed approach consists of iteration through three components: transcriptome measurements, Network Component Analysis (NCA), and chemoinfomatic analysis. The investigation starts from micro array experiments of wild-type E. coli response to RNOS and the initial transcription factor (TF)-promoter connectivity from the literature database, which is incomplete for the conditions of interest. NCA first deduces the activities of TF upon RNOS challenge based on wild-type micro array data. The activities of TFs, rather than the expression levels, provide critical information for deducing chemical mechanisms and signaling pathways. Among the TFs perturbed, chemoinfomatic analysis allows the classification of affected TFs into direct and indirect targets and the identification of potential chemical mechanisms for signaling. The perturbed TFs are then deleted using chromosomal knockout techniques. The resulting knockout strains are further tested for responses to RNOS and the transcriptome data are analyzed in NCA to deduce revised target TFs and transcription networks. If hew TF targets are predicted, then the iteration continues. Otherwise, the predicted networks are verified using genetic and biochemical experiments. [unreadable] [unreadable] [unreadable]